← World of AI
Neural Networks
Deep Belief Network
Stacked probabilistic layers trained greedily — a stepping stone to modern deep learning.
Built by training restricted Boltzmann machines one layer at a time, then fine-tuning the stack. In the mid-2000s this was the practical answer to the problem that deep networks would not train.
Better activations, initialisation schemes and far more data made the greedy pre-training step unnecessary, but the architecture is a fair marker of where deep learning restarted.
Also in Neural Networks
JOIN NOW
Begin the first module
It is free, it is the real curriculum, and if it is not for you, you have lost nothing but an evening.
Join any time · Build AI skills at your pace